This paper introduces a new learning algorithm for artificial neural networks, based on a fuzzy inference system ANBLIR.It is a computationally effective neuro-fuzzy system with parametrized fuzzy sets in the consequent parts of fuzzy if-thenrules, which uses a conjunctive as well as a logical interpretation of those rules. In the original approach, the estimationof unknown system parameters was made by means of a combination of both gradient and least-squares methods. Thenovelty of the learning algorithm consists in the application of a deterministic annealing optimization method. It leads to animprovement in the neuro-fuzzy modelling performance. To show the validity of the introduced method, two examples ofapplication concerning chaotic time series prediction and system identification problems are provided.
Sep 7, 2021
Aug 9, 2018
|Neuro-fuzzy modelling based on a deterministic annealing approach||Sep 7, 2021|
Czabański, Robert Korbicz, Józef - red. Uciński, Dariusz - red.
Gierlak, Piotr Muszyńska, Magdalena Żylski, Wiesław (1938-2018) Jurczak, Paweł - red.
Łęski, Jacek M. Henzel, Norbert Rutkowska, Danuta - ed. Zadeh, Lotfi A. - ed.
Rutkowska, Danuta Nowicki, Robert K. Rutkowska, Danuta - ed. Zadeh, Lotfi A. - ed.
Patton, Ron J. Lopez-Toribio, Carlos J. Uppal, Faisel J. Korbicz, Józef - red. Patton, Ronald J. - red.
Łęski, Jacek M. Czogała, Tomasz Korbicz, Józef - red. Uciński, Dariusz - red.
Łęski, Jacek M. Rutkowska, Danuta - ed. Kacprzyk, Janusz - ed. Zadeh, Lotfi A. - ed.
Tatjewski, Piotr Ławryńczuk, Maciej Korbicz, Józef - red.